Ever since Microsoft released a bunch of details about Windows 8 at their BUILD 2011 conference a few weeks ago, I've been getting questions like, "Is it true that Microsoft is finally getting serious about desktop virtualization with Windows 8?" or "Will Windows 8 go after Citrix for desktop virtualization?"
The answer is "no." Here are a bunch of reasons why not:
Windows 8's "desktop virtualization" support is really just the addition of a client hypervisor, not some "real" effort
The confusion stems from the fact that the terms "virtual desktops" and "desktop virtualization" mean different things to different people. So when we see articles like CRN's "Windows 8 to Support Desktop Virtualization" or CTOEdge's "Windows 8 and New Desktop Virtualization Realities." it's easy to see how people might misconstrue that to mean that Microsoft is somehow extending Windows 8 to compete against Citrix XenDesktop or VMware View. But in reality, both articles (and others like them) are really just talking about Windows 8 having an option for a client hypervisor built-in, which is nice, but not a game changing desktop virtualization push from Microsoft. (More details and a discussion about the Windows 8 client hypervisor can be found in Gabe's article from last week.)
Microsoft doesn't know their ass from a hole in the ground when it comes to desktop virtualization
Have you seen Microsoft's VDI, err, Remote Desktop Virtualization Host offerings? Spend ten minutes playing with that and you'll see why Microsoft is quick to shove as many customers to Citrix XenDesktop or Quest vWorkspace as possible.
We've spent probably 50k words on this website writing about the future of the desktop, Microsoft's role, etc., so I don't have to repeat too much. But let me just say that Microsoft hasn't exactly been the "enlightened leader" when it comes to desktop virtualization. Hell, you could argue that they're doing everything they can to *prevent* desktop virtualization from catching on. (VECD, App-V in SA, licensing by device instead of user, no layering built-in to windows, old school profiles still in Win8…)
Let's be real. What are the chances that Microsoft is going to (a) suddenly decide to focus on desktop virtualization in Win8, and (b) have that hastily-decided focus be the right focus, and (c) actually be able to execute on that? I think there's a better chance that a satellite falls on my head this weekend.
Microsoft is a big ship with a tiny rudder
Let's go back to that imagined scenario where Microsoft does suddenly become enlightened about desktop virtualization. Even if they had a great strategy today, what's their launch vehicle for that? Windows 8? What's that...at least a year away? Even if that's true, don't you think CitrixVMwareQuestVirtualBridges will have a new version or two by then? Microsoft doesn't exactly have a good track record in the "oh sh*t we're really far behind so let's throw a lot of resources at the problem to win" department. The path for Windows 8 is paved with tablets, not virtual desktops.
Windows 8 for tablets will change the game
Oh wait, no it won't. Sure, it will have a great UI. Metro is cool, etc. The problem is that in the "Post PC" world, the IT-supported Windows platform apps are going to be the old shitty ones that we can't rewrite for the cloud/SaaS/HTML5/iOS/Android/RIA. And do you think those old shitty apps are going to magically get a Metro UI? (This is a great opp for a software vendor btw.. kind of like Citrix Project Vertigo for the 21st century.)
And don't be fooled by promises of "Metro Style Remote Desktop App." That just means that you'll be able to launch a remote desktop session via a Metro RDC client with your multiple greasy fingers and no keyboard. But what happens when you actually connect to that remote desktop and you find that you have fat fingers and no keyboard? Pan and zoom baby!! Pan and zoom… (Oh yeah, sure, if your remote Windows desktop is running Metro and you have a Metro app running on it, then RemoteFX and RDP will also be able to remote multitouch. But what's the use case for that really? It seems that if someone took the time to rewrite their Windows app for Metro, then wouldn't you just compile it for ARM and run it on your tablet locally? Or better yet, since you're rewriting the thing, just rewrite it for an app platform you don't have to remote, like something that works with any HTML5 browser.)
BTW, keep in mind that no matter how serious Microsoft gets about Windows 8 for ARM for tablets, remember that you won't be able to run x86 Windows apps on Win8 for ARM. So really Win8 with Metro and/or ARM is only "Windows" in the sense that it will be easy-ish for current Windows developers to write for it. It's not "Windows" in the sense of "let's make all our old Windows crap work."
Windows 8 will f*ck up our Windows 7 migration plans
Ok, this isn't technically a statement about Windows 8's desktop virtualization capabilities, but it's something to think about. If Windows 8 really does come out in 2012, that's still more than a year before XP support expires, so maybe some of us can wait and just jump from XP to 8?
The good news is that in terms of core architecture, Windows 7 and Windows 8 are very similar. You can still run Windows 8 on x86. If you have a keyboard and mouse then you don't have to run Metro. And it seems that the jump from 7 to 8 will be more like an "in place upgrade" as opposed to an actual migration. So at this point I'm cool with everyone going to Windows 7 for everything, and when Windows 8 comes out, we'll all go there. But when it comes to desktop virtualization, we're going to use Windows 8 in the same way we're using Windows 7. Nothing new here.